The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions
The sloping mire landscape of the investigation area, in the southern Andes of Ecuador, is dominated by stagnic soils with thick organic layers. The recursive partitioning algorithm Random Forest was used to predict the spatial water stagnation pattern and the thickness of the organic layer from ter...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2014-01-01
|
| Series: | Applied and Environmental Soil Science |
| Online Access: | http://dx.doi.org/10.1155/2014/603132 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849691446839345152 |
|---|---|
| author | Mareike Ließ Martin Hitziger Bernd Huwe |
| author_facet | Mareike Ließ Martin Hitziger Bernd Huwe |
| author_sort | Mareike Ließ |
| collection | DOAJ |
| description | The sloping mire landscape of the investigation area, in the southern Andes of Ecuador, is dominated by stagnic soils with thick organic layers. The recursive partitioning algorithm Random Forest was used to predict the spatial water stagnation pattern and the thickness of the organic layer from terrain attributes. Terrain smoothing from 10 to 30 m raster resolution was applied in order to obtain the best possible model. For the same purpose, several model tuning parameters were tested and a prepredictor selection with the R-package Boruta was applied. Model versions were evaluated and compared by 100 repetitions of the calculation of the residual mean square error of a five-fold cross-validation. Position specific density functions of the predicted soil parameters were then used to display prediction uncertainty. Prepredictor selection and tuning of the Random Forest algorithm in some cases resulted in an improved model performance. We therefore recommend testing prepredictor selection and tuning to make sure that the best possible model is chosen. This needs particular emphasis in complex tropical mountain soil-landscapes which provide a real challenge to any soil mapping approach but where Random Forest has proven to be successful due to the testing of model tuning and prepredictor selection. |
| format | Article |
| id | doaj-art-e8acb4462263485086e838db9f41ca51 |
| institution | DOAJ |
| issn | 1687-7667 1687-7675 |
| language | English |
| publishDate | 2014-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Applied and Environmental Soil Science |
| spelling | doaj-art-e8acb4462263485086e838db9f41ca512025-08-20T03:21:02ZengWileyApplied and Environmental Soil Science1687-76671687-76752014-01-01201410.1155/2014/603132603132The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest PredictionsMareike Ließ0Martin Hitziger1Bernd Huwe2Department of Geosciences/Soil Physics Division, University of Bayreuth, Universitaetsstraße 30, 95447 Bayreuth, GermanyETH Zürich, Environmental Natural and Social Sciences, Universitaetsstraße 22, 8092 Zürich, SwitzerlandDepartment of Geosciences/Soil Physics Division, University of Bayreuth, Universitaetsstraße 30, 95447 Bayreuth, GermanyThe sloping mire landscape of the investigation area, in the southern Andes of Ecuador, is dominated by stagnic soils with thick organic layers. The recursive partitioning algorithm Random Forest was used to predict the spatial water stagnation pattern and the thickness of the organic layer from terrain attributes. Terrain smoothing from 10 to 30 m raster resolution was applied in order to obtain the best possible model. For the same purpose, several model tuning parameters were tested and a prepredictor selection with the R-package Boruta was applied. Model versions were evaluated and compared by 100 repetitions of the calculation of the residual mean square error of a five-fold cross-validation. Position specific density functions of the predicted soil parameters were then used to display prediction uncertainty. Prepredictor selection and tuning of the Random Forest algorithm in some cases resulted in an improved model performance. We therefore recommend testing prepredictor selection and tuning to make sure that the best possible model is chosen. This needs particular emphasis in complex tropical mountain soil-landscapes which provide a real challenge to any soil mapping approach but where Random Forest has proven to be successful due to the testing of model tuning and prepredictor selection.http://dx.doi.org/10.1155/2014/603132 |
| spellingShingle | Mareike Ließ Martin Hitziger Bernd Huwe The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions Applied and Environmental Soil Science |
| title | The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions |
| title_full | The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions |
| title_fullStr | The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions |
| title_full_unstemmed | The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions |
| title_short | The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions |
| title_sort | sloping mire soil landscape of southern ecuador influence of predictor resolution and model tuning on random forest predictions |
| url | http://dx.doi.org/10.1155/2014/603132 |
| work_keys_str_mv | AT mareikeließ theslopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions AT martinhitziger theslopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions AT berndhuwe theslopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions AT mareikeließ slopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions AT martinhitziger slopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions AT berndhuwe slopingmiresoillandscapeofsouthernecuadorinfluenceofpredictorresolutionandmodeltuningonrandomforestpredictions |